When data changes, it gets corrected - and the old version is usually thrown away. But what if there was something in that record which could be valuable for your targeting? David Reed finds out how data quality is evolving into data intelligence.
Data changes. It is the unavoidable fact of data management that you can spend all day improving the quality of your records, validating addresses and getting it fit for purpose, then the moment you leave for home, somebody moves house, dies, marries, gets gender reassigned or just a new email address.
It is the reason why there is a £22.8 million market in suppression files. Identifying deceaseds, goneaways and changes of address is central to the whole process of data hygiene. It is barely two decades since matching to such files was an innovative step. Since then, the coverage and quality of suppression files has improved, but the essential practice has stayed the same.
While improvements to these files continue, there is a new approach emerging which goes beyond just flagging or removing files where something has changed. Instead, there is a recognition that these changes are intelligence in their own right that tells you something about the subject, their address and your database as a whole.
“Recently we have been looking at historical segmentation of goneaways, where they sit in different segments of the database and the potential of using that for insight,” says Rupert Tomalin, bureau director at meta-morphix. “Where somebody who is a high-value customer is flagged, that tells you there is a need to acquire more like that.”
Recognising this link between the negative targeting of suppression and the positive targeting of acquisition has led to the creation of Qinetic. It contains relocated addresses, new occupier data, deceaseds and also geo-demographic profiles, dates of birth and financial risk indicators.
When a database is matched against Qinetic, the old record is not discarded. Instead, data on the new occupier is provided as well as the relocated address for the old occupier. Geodemographic profiles allow for comparisons between previous and existing residents to determine whether the address is still the right type of prospect.
“The key is that you are not suppressing a record. You flag them as a goneaway so you can reapply them at a later data, because it can take a while for somebody to appear at a new address,” says Tomalin. With conventional suppression procedures, there is always a gap while the moving data is “in flight” - that is, between the time when a data owner picks up that they have moved and when they are confirmed at a new address.
This is one reason why matches to changes of address are typically less than half those to goneaways (0.77 per cent compared to 1.62 per cent according to the DataIQ Suppression Files Report 2011). Tracking customers has not been common practice because the moved-to address element is not always available at the point when a customer is identified as having moved away from their current address. (Since only half of home-movers use a redirection service, there is also a gap in coverage.)
“If you flag that person as goneaway, finding them again is more valuable to you than getting a new customer onboard,” points out Tomalin. It may cost 30p to append a new address for a customer who has moved, but if they are, for example, a high value charity donor giving more than £20 per month, maintaining that contact makes clear commercial sense.
The database has been built to capture these internal dynamics and fill the gap which redirection data leaves. Maintaining an address history for individuals can help to avoid the problem of losing customers and assuming they have just lapsed.
Tomalin points out that it also opens up another opportunity through using the geodemographic codes on Qinetic to compare old and new circumstances. Areas and streets can go up and down in terms of their social status and the income levels of the people living there. Trading up or down the property ladder releases equity or diminishes disposable income, changing the ability of an individual to purchase products and services.
That is a valuable piece of intelligence which can be gained from understanding the internal dynamics of the database. It can also provide indications of new opportunities - for example, owners who recently installed a conservatory but have now moved house are likely to have taken the furniture with them. Promoting appropriate fittings to the new owners is “a new angle”, he notes.
Data intelligence is not just generated - and lost - in the process of cleaning. There are also potentially valuable insights to be gained from tracking data as it enters via different feeds or is migrated out of one system into another. In the push towards creating a single view of the customer, indirect variables are missed that could provide useful marketing fuel.
“Data gets provided into the marketing environment through multiple channels, such as online, call centre, or campaign responses. The usage each customer makes of those touchpoints tends to get trimmed down when the data is fed into a SCV unless you have really thought it through and specified it should be kept,” says Nigel Magson, managing director of Adroit.
He identifies marketing services providers, who are typically tasked with the job of creating the SCV, as being at the root of this loss of intelligence. “They have a data structure that they fit every client’s data to. In the process, stuff goes missing,” he warns.
Address history is one critical piece of intelligence that is typically discarded during this process of data integration. “We do a fair amount of legacy marketing, for example with insurance products trying to understand the profile of customers at fixed points in time. You need to go back to the address they lived at when you acquired them - there is no point profiling people on where they live now if you are looking for similar prospects who are in the market,” he says.
The reason why a customer purchases insurance usually relates to their circumstances at the time, such as moving to a new house, buying a new car or having children. Insight on those drivers needs to be tied to where they were at the time, but the constant integration of data and migration between systems means previous addresses are often discarded.
Adroit uses a data fact-finder template which maps the history of customer data. “We often find out there has been a systems change or that proper communications history has not been kept. For one publishing, his agency built a reporting engine from this history that allowed it to tie sales back to prospect marketing activity. It also tracked 25 campaign variables and condensed 97 profiling variables.
All of this data is pulled out of the database and filtered down into an Excel spreadsheet. “It means we can aggregate cells across multiple campaigns and get the significance of what we are looking at. Even if the client is losing data or not pushing it through, we can at least see what is significant,” says Magson.
He believes that part of the blame for lost data intelligence has to be placed on external suppliers. “Agencies are often not doing analysis linked to communications history. They are going into the transactional file looking at sales data, but does that uplift come from specific contacts?” he asks.
That is the ultimate question for all marketing. Indeed, analytics is the pursuit of the reasons why behaviours happen, especially purchasing. What is just being recognised is that the dynamics of data which are usually only considered from a data quality perspective could also prove to be an important dimension in insight.